A Qualitative Study to Optimize a Community-Based Pediatric TB/HIV Prevention Intervention and Implementation Strategy

Tuberculosis (TB) remains a top 10 cause of child mortality in sub-Saharan Africa. Short-course combination TB preventive therapy (TPT) regimens are highly effective with high completion rates and therefore hold promise to reduce the 1 million pediatric TB patients and 233,000 child TB deaths annually. TPT effectiveness requires improvement in access to pediatric TB preventive care. The World Health Organization (WHO) estimates only 23% of the 1.3 million TB-exposed children under 5 years old initiated TPT in 2017. The investigators' previous data suggest 50-75% of TB-exposed children under 5 years old fail to be either identified or linked to TB preventive care. Children under 5 years are at exceptionally high risk of TB disease due to not only immature immune systems, but also HIV infection and HIV exposure. Guidelines have long recommended TPT for children under 5 years, but have recently been expanded to children under 15 years. Developing and testing innovative health care delivery models to accompany new WHO-endorsed short-course combination TPT regimens is essential to reduce TB-associated global child morbidity and mortality.

Integrated community case management (iCCM) is a cost-effective intervention that increases child survival by bringing curative therapies to at risk children in the community. Recently, algorithms for screening pediatric patients with household exposure to TB have been simplified to include symptom screening only, eliminating costly laboratory and radiographic evaluations that are barriers to treatment and do not improve sensitivity and specificity of the assessment. The CHIP-TB Trial plans to assess whether integrated community-based pediatric TB/HIV prevention services is feasible, acceptable and effective in increasing TPT uptake among child TB contacts under 15 years in a cluster-randomized trial in South Africa and Ethiopia, two high-burden, resource-limited settings.

Nested within this larger trial, the investigators will use qualitative research to (1) describe the social context in which the community health teams perform the team's duties, (2) make context-specific adaptations to the intervention to maximize its effectiveness, (3) define a context-specific implementation strategy and evaluation plan, and (4) learn important lessons for the future scale up and dissemination of the community-based intervention.

The investigators will conduct 10-15 paired semi-structured individual in-depth interviews (IDI) for each of the four key stakeholder groups: (1) caregivers of TB-exposed children, (2) community health team members, (3) clinic-based nurse and clinic or hospital-based physician providers, and (4) health system managers and policy makers, for a total of ~120 IDIs (~60 in South Africa and ~60 in Ethiopia) both pre and post trial. Before the intervention the investigators will conduct a one-hour interview that will explore the inner setting of the community health team's environment from each stakeholder perspective and will develop and decide upon key intervention adaptations and implementation strategies for the community-based TB/HIV prevention program that will be assessed with a one-year cluster-randomized trial. Upon trial completion, the investigators will conduct another 10-15 semi-structured IDIs with each stakeholder group in each country (total ~120 IDIs) to identify key facilitators and barriers to the implementation. Thematic content analysis with an inductive approach will be used to analyze transcripts for both a priori and emergent domains of interest. A coding scheme will be adapted from Consolidated Framework for Implementation Research (CFIR) constructs. Free listing, ranking, and cultural consensus will be used to identify intervention adaptations and implementation strategies that are context specific. The investigators will use Smith's Saliency Index to analyze the free listed data, mean rank order to assess ranking exercises, and the Eigenvalue to evaluate consensus across groups all using the ANTHROPAC ® 4.0 software.